15:04:00 #startmeeting kde-sig -- http://fedoraproject.org/wiki/SIGs/KDE/Meetings/2012-04-17 15:04:00 Meeting started Tue Apr 17 15:04:00 2012 UTC. The chair is jreznik. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:04:00 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:04:07 #meetingname kde-sig 15:04:08 The meeting name has been set to 'kde-sig' 15:04:11 * jreznik is updating wiki 15:04:17 #topic Roll Call 15:04:24 * rnovacek is here 15:05:03 Present. 15:06:00 here (doing a bad job of multitasking, and good job of distracting myself today) 15:06:24 * than is present 15:06:25 * jreznik is sickday today, still not feeling well but everything for meeting :) 15:06:36 * ltinkl is present 15:06:52 #chair jreznik ltinkl Kevin_Kofler rdieter than rnovacek 15:06:52 Current chairs: Kevin_Kofler jreznik ltinkl rdieter rnovacek than 15:07:01 #info Kevin_Kofler jreznik ltinkl rdieter rnovacek than present 15:07:15 #topic Agenda 15:07:43 ideas - test day wrap-up... 15:09:05 anything for agenda? 15:10:15 perhaps blocker bugs for f17 if we have 15:11:09 Oh, and what about shipping the polkit-kde authorization editor KCM? I noticed that some other distros are already shipping it. 15:11:24 It's the only GUI to edit PolicyKit 1 authorizations. 15:11:31 (There's not even a GNOME one.) 15:11:55 was it ever really! finished? 15:12:03 Kevin_Kofler: i guess that would mean we should package it first. 15:12:08 than: I'm not aware of any 15:13:05 https://projects.kde.org/projects/extragear/base/polkit-kde-kcmodules-1/repository 15:13:06 speaking of stuff never finished, I just updated kiosktool from a git snapshot, it supports kde4 now, and (reportedly) actually works now. 15:13:23 Thomas Balling Sørensen did some stuff on that in September 2011. 15:13:57 looks like more informal meeting today without agenda :) 15:13:59 I don't know whether it's good enough to ship though. I only know that some distros are shipping it, but for all I know they might be shipping broken crap. ;-) 15:15:00 rdieter: Good news, is that in Fedora proper yet? Or only in kde-redhat? 15:15:11 Kevin_Kofler: imo we need to check it before, just avoiding to add broken stuff 15:15:19 hrm, I updated fedora git, but it may have been half-retired or blocked or something 15:15:24 my koji build didn't go through 15:16:02 Re blocker bugs, I nominated the purple mess for final blocker status (as one argument for not having it block the Beta was "it's only a Beta", which obviously doesn't apply to the Final). 15:16:18 No decision so far, I guess they'll only seriously look at the proposed blockers once we enter the Final Freeze. 15:16:21 what's 'the purpose mess'? 15:16:26 the purple mess even 15:16:41 The cirrus 16 bpp issue. 15:17:08 oh ok (though it was some reference to libpurple... :) ) 15:18:38 There's also the livecd-tools size regression which makes us fail the blocking size criterion, a fix for that is already in Bodhi, so hopefully that'll be sorted out soon. 15:18:45 ok, confirmed kiosktool is blocked, and no < master/f18 branch exists 15:19:07 (It turned out they accidentally and unnecessarily put multiple copies of the kernel on the ISO, which is what bloated it so much.) 15:19:16 rdieter: one of my students will work on extending kiosktool 15:19:17 ugh, that may mean a re-review, but meh. I'll investigate more after meeting 15:19:21 ltinkl: yay! 15:21:58 jreznik: ok, let's move on to your test day wrap up topic 15:22:46 And thanks to rdieter and jreznik for getting qt-at-spi into Fedora, another wanted package that is in now. 15:23:51 Kevin_Kofler: the idea behind qt-at-spi is to be able to do automatic test via dogtail 15:24:11 and it's still crashing heavily (even 0.3.1) as reported by vhumpa 15:24:27 also the idea is to provide kcm to enable it easily 15:25:48 qt-at-spi is also useful for people who need the Orca screen reader. But automatic testing is also very nice. Hopefully the crashes can be sorted out. 15:26:11 Kevin_Kofler: of course it's great for people who need screen readers 15:26:27 #topic Test Day wrap-up 15:26:42 I guess Dogtail could also be abused to automate GUI apps in some cases. :-) 15:27:02 #link https://fedoraproject.org/wiki/Test_Day:2012-04-10_KDE_4.8 15:27:02 (GUI scripting) 15:27:38 we have received quite a lot of test results 15:27:48 got a lot of good testing, feedback, and bugs. 15:28:20 I think it was worth the work, next time we will need more time for creating better test cases 15:28:40 but still we have more confidence in F17 is at least booting :) 15:28:56 Also a couple very strange bug reports, but most of the reports seem to make sense. Now I wonder whether we can actually fix them or get them fixed by upstream… 15:29:20 Kevin_Kofler: I'd like to go through tmrw (same for desktop validation tests ones) 15:29:59 * rdieter tried triaging some of them at least, but indeed, most need more love (confirming, re-testing, sending upstream where appropriate) 15:29:59 we should try to upstream it but also we should push on it to get fixes 15:30:16 rdieter: thanks, I saw your feedback 15:30:31 it would be great to offer some feedback to our users as thank you for testing 15:30:50 so I'd like to ask you to love these bugs :) and for me it's command :) 15:32:24 indeed, most of these (and other recent) bugs were very good ones, so explictly offering them thanks is the least we can do. and, I will say so next chance I get to comment in any of those bz's 15:35:07 anything else for test day? 15:35:19 jreznik: +1, i will go througg bugzilla and add some comments 15:39:40 Hmmm, one more thing: What about the firstboot theming issue? 15:40:12 #topic Firstboot theming 15:40:22 Kevin_Kofler: good topic 15:40:36 if we will prepare patch, it will be merged into firstboot 15:40:38 I have a deal 15:40:40 Has anybody looked at making firstboot detect the KDE spin and run xsettings-kde there as we discussed? If not, should we restore the gtkrc hack which makes oxygen-gtk the default theme for root on the KDE spin? 15:41:11 So I have an idea how to fix firstboot (see above), but so far I haven't been able to look into implementing it. 15:41:18 so if you have any idea how to detect it, I can take a look 15:41:42 i'd guess the path of least-resistance would be a gtkrc 15:42:08 (that'll only work as long as firstboot is gtk2-based of course) 15:42:12 I'd say a good and on-topic method would be a test like "exists(oxygen-gtk2) and not exists(adwaita-gtk2)". 15:42:52 rdieter: It's also possible for gtk3. The drawback is the same in both cases, it forces the theme for root even in other desktops. 15:43:21 oh right, I was thinking GTK2_RC_FILES 15:43:37 your proposal is better 15:47:14 Kevin_Kofler: so your 'if...' pseudo code would conditionally run xsettings-kde ? 15:48:52 Yes. 15:49:11 Probably also testing that /usr/bin/xsettings-kde exists before trying to run it would be a good idea. :-) 15:51:00 Though actually I'm not sure it would be sufficient… How do we set the default theme for kcm-gtk and xsettings-kde to use? There's some script in /etc/kde/env, isn't there? That's not going to be run in firstboot. :-/ 15:51:57 (unless we run it explicitly) 15:53:00 well, you'd have to test it all, sure. :) 15:53:46 but sure, since xsettings-kde currently groks it from .gtkrc2-kde4 or whatever, right? 15:54:15 * rdieter checked, ~/.gtkrc-2.0-kde4 is the one 15:54:20 Right, and that's copied by some /etc/kde/env script, what a mess. 15:54:44 https://fedorahosted.org/kde-settings/browser/trunk/etc/kde/env/gtk2_rc_files.sh 15:55:02 ok, firstboot isn't a user session, so won't get any of that by default alright 15:55:05 So the firstboot hack would have to invoke that before invoking xsettings-kde. 15:55:12 yup 15:55:45 Unless we do the copy in the live kickstart. 15:55:48 (for root) 15:56:09 uf 15:56:36 the live kickstart would be a nice addition, but wouldn't help non-live installs 15:56:46 DVD, netinstalls, yada yada 15:56:47 Another big mess is that this also sets GTK2_RC_FILES; without that set, I don't think xsettings-kde will find the file. 15:56:56 rdieter: The whole hack is live-CD-specific. 15:57:15 (live-install-specific, to be more precise) 15:57:33 firstboot from DVD install works ok ? 15:57:47 Well, I guess it depends on what you checked. 15:58:04 If you manage to install without adwaita, you'll also get the ugly fallback theme there, otherwise you'll get adwaita. 15:58:08 ok, so I was hoping we have something more robust to handle both(all!) cases 15:58:10 This has always been like that. 15:58:44 but baby steps, if getting only kde-live working is a first step, so be it 15:59:30 Uhm, what if, instead of invoking xsettings-kde, if we have Oxygen and no Adwaita, we write a gtkrc2-firstboot which sets the theme to Oxygen and set GTK2_RC_FILES to that. 15:59:34 Do it all inside firstboot. 16:00:00 That will do the right thing whenever you install with the KDE group and none of the comps groups with Adwaita in them. 16:00:04 right that was my first suggestion about 15 minutes ago. :) 16:00:10 (and if you do have Adwaita, it will use Adwaita) 16:00:41 well, the basic idea, anyway, works for me. 16:02:08 Kevin_Kofler: if you want to work on it, i'll defer to your perference (ie, the one doing the work gets more say in the matter). :) 16:02:26 we are out of time -> #fedora-kde 16:02:35 nod 16:02:41 #endmeeting